Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 23 Sep 2015 15:27:02 -0700
From:      Tim Gustafson <tjg@ucsc.edu>
To:        Ryan Stone <rysto32@gmail.com>
Cc:        Michael Fuckner <michael@fuckner.net>, "freebsd-fs@freebsd.org" <freebsd-fs@freebsd.org>
Subject:   Re: Switching from MFI to MRSAS
Message-ID:  <CAPyBAS6U_ObeVCQS1b15B3iT1nR%2BHwsA1r72hjgNtRsgOkTuTg@mail.gmail.com>
In-Reply-To: <CAFMmRNwMv6ZAWhyBiSwufh1cCxxBYz5aOt4s1_0_MvOR8JZrMA@mail.gmail.com>
References:  <CAPyBAS4-oRjrybxu81nPO1Xb0WPdYzJRoEZ3t8zjNUrxFsiZpw@mail.gmail.com> <56023C15.7080007@fuckner.net> <CAPyBAS6p_=Ld6yiis00L3GxCFgq7JwUXkvg=mzni3ZHHEC1Z3Q@mail.gmail.com> <5602D9E8.2090903@fuckner.net> <CAPyBAS5TyBYjOp6wRSkpOy35m%2BcQVLn3MNFS3=y2wWYtpjobQw@mail.gmail.com> <CAFMmRNwMv6ZAWhyBiSwufh1cCxxBYz5aOt4s1_0_MvOR8JZrMA@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
> The mrsas driver is interesting.  Your RAID volumes show up as da0, da1,
> etc, but they really shouldn't.  I'm doubtful that TRIM will work here
> because the drives are not truly directly attached, even though the mrsas
> driver is misleading the kernel into thinking that they are.
>
> You can always give a try, though.

Just to be clear: we have the PERC controller configured in
"pass-through" mode, which I understand to mean "HBA" mode.  We are
not using any of the RAID features of the PERC card.  Instead, the
server is configured as ZFS-on-root in a RAID 0+1 configuration:

  pool: tank
 state: ONLINE
  scan: scrub repaired 14.5K in 0h13m with 0 errors on Thu Sep 17 20:47:17 2015
config:

NAME             STATE     READ WRITE CKSUM
tank           ONLINE       0     0     0
 mirror-0       ONLINE       0     0     0
   mfisyspd0p3  ONLINE       0     0     0
   mfisyspd1p3  ONLINE       0     0     0
 mirror-1       ONLINE       0     0     0
   mfisyspd3p3  ONLINE       0     0     0
   mfisyspd2p3  ONLINE       0     0     0

where mfisyspd0, mfisyspd1, mfisyspd2 and mfisyspd3 are Samsung SSDs,
not RAID volumes.

So in this case, I believe it would be "correct" for the RAID
controller to report the devices as "da0", "da1" and so on.

-- 

Tim Gustafson
tjg@ucsc.edu
831-459-5354
Baskin Engineering, Room 313A



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAPyBAS6U_ObeVCQS1b15B3iT1nR%2BHwsA1r72hjgNtRsgOkTuTg>